Enterprise search vendors approached the problem of search by assuming all data is created equal. They have attempted to provide a single solution to all search needs of the enterprise. This ‘one size fits all’ approach is flawed when you consider the complexity of large sets of both structured and unstructured data. These vendors have tried to mimic what Web search engines like Google and Bing have done for the Internet. Web search works to a large extent because you are expected to contextualize the search results yourself, and they are evolving to provide more context for their search results. Lack of contextualized search has been the biggest shortcoming of Enterprise Search. Enterprise users expect their search results to present within the context of what they are attempting to search. Capturing the true intent of the search presents a huge challenge and part of the problem is “all data is NOT created equal.”
Navigating through the complex maze of both current and historical data requires a new approach to search. This is where purpose-built search applications have been successful because they understand the data as well as the users from a particular domain. When you contextualize the results, the users can easily navigate through this maze and get their questions answered much quicker than having to stitch together different bits of information from multiple search results. Moreover, the system can start telling the user rather than the user asking the system. There are a number of technical challenges around this, and a true intelligent search engine has been the Holy Grail of Enterprise Search.
So what does it mean by contextualizing search? To answer this question let’s look at Google Search. In the early days when you searched for ‘AUS to SFO’, the results were much different than they are today. In the old days the results were based on the contents of your search string, not its meaning. Today the engine understands that AUS and SFO are two cities and makes a best guess that your purpose is to find an airplane ticket for an upcoming trip. When it comes to product and component data, let’s say you are searching for ‘RES 10K 10% SMT .’ If the engine takes just the contents of the search string, the results may be closer to what you are looking for as long as these contents appear somewhere on the part record (in most cases they are not). However, a true intelligent search engine that understands the product data domain can infer that you are looking for a ‘Resistor (type) with 10K resistance, which has a 10% tolerance and can be surface mounted (SMT)’. Here the true intent if the user’s search is captured in the search results. This is what purpose-built search applications are attempting to do and have made long strides in this area.
Modern enterprises require their knowledge workers to be more agile as the products they build are getting more complex and have shorter cycles as their competitors are fighting tooth-and-nail to outdo them. Also, for their knowledge workers to be more efficient, they are leaving no stone un-turned to provide access to their data to help them make better and faster decisions. New breeds of search platforms like Encompass can provide the competitive edge these modern enterprises are striving for. Whether it is your part data, BOM data, current or historical, and whether your users are design engineers, product operations, manufacturing or product support, Encompass provides a suite of applications that understand your data, as well as your users. I am not saying Encompass has found the Holy Grail, but I do believe it has found the trails that lead to it. Navigating through your entire product data hasn’t been this easy before!